Journal of Human Hypertension (1999) 13, 787–791
 1999 Stockton Press. All rights reserved 0950-9240/99 $15.00
ORIGINAL ARTICLE
Effectiveness and metabolic effects of
perindopril and diuretics combination in
primary hypertension
MS Elisaf, J Theodorou, H Pappas, N Papagalanis, K Katopodis, R Kalaitzidis and
KC Siamopoulos
Department of Internal Medicine, University of Ioannina Medical School, Greece
The effectiveness as well as the metabolic effects of the
combination of diuretics [hydrochlorothiazide (HCT) vs
indapamide (IND)] and perindopril (P) in 14 patients (7
male, 7 female) aged 37–62 years with mild idiopathic
hypertension were studied. Following a 4-week wash-
out period and a 4-week period of monotherapy with P
(4 mg/daily), IND (2.5 mg/daily) or HCT (25 mg/daily) was
added for 4 weeks. Selection of the diuretic agent was
random. Following a 4-week wash-out period from the
diuretic, in which only P was given, the alternative
diuretic was administered for another period of 4 weeks.
P decreased blood pressure levels significantly. How-
ever, the drug was more efficacious in patients with
higher plasma renin activity (PRA). Combination treat-
ment induced an additional decrease in the blood press-
ure levels, mainly in patients with lower PRA. The com-
bination of P ⴙ HCT was more effective than the
combination P ⴙ IND. The addition of either HCT or IND
Keywords: perindopril; hydrochlorothiazide; indapamide; plasma renin activity; combination drug therapy for hypertension;
metabolic effects of antihypertensive treatment
Introduction
Angiotensin-converting enzyme (ACE) inhibitors
comprise a class of antihypertensive drugs with a
very low side-effect profile.1 However, single drug
therapy is effective in approximately half of hyper-
tensive patients. In the remaining patients, a syner-
gistic effect upon hypotensive efficacy has been
observed when even small doses of diuretics are
added.2–5 Even though diuretics exert adverse meta-
bolic effects, ACE inhibitors appear to exhibit a neu-
tral or even a beneficial effect on lipid and carbo-
hydrate metabolism, leading to an attenuation of the
biochemical sequalae commonly associated with
diuretic therapy.6–11 However, indapamide has been
reported to induce a moderate change in the serum
metabolic parameters.12,13
Therefore, we undertook the present study to
examine the efficacy as well as the metabolic effects
Correspondence: Dr Moses S Elisaf, Associate Professor of Medi-
cine, Department of Internal Medicine, University of Ioannina,
Medical School, GR 451 10 Ioannina, Greece
Received 10 January 1999; accepted 25 February 1999
http://www.stockton-press.co.uk/jhh
evoked a small but statistically significant increase in
serum glucose levels while fasting as well as during the
75 g oral glucose challenge. However, insulin levels did
not change significantly during the study. Small but not
statistically significant changes in serum electrolytes
and lipid parameters were observed during the various
phases of the study, while a statistically significant
increase in the serum uric acid was noticed when the
combination P ⴙ HCT was given.
We conclude: (1) P in small doses is an effective and
safe antihypertensive agent, (2) PRA has a predictive
value in determining the effectiveness of P treatment,
(3) the combination of P with small doses of HCT or IND
is more efficacious than P alone, (4) the combination
treatment has adverse effects in the carbohydrate toler-
ance, while there are not significant changes in serum
electrolyte and lipid parameters.
of the combination of indapamide or hydrochloro-
thiazide with perindopril.
Materials and methods
A total of 14 patients (7 male, 7 female) aged 37–62
years with mild to moderate primary hypertension
(sitting diastolic blood pressure (BP) ⭓95 mm Hg
and ⭐114 mm Hg) were studied. All patients had
been previously identified by the investigators as
requiring a drug combination to control their BP.
Criteria for exclusion were significant cardiac, hep-
atic, renal, and thyroid diseases, diabetes mellitus,
and weight in excess of 20% of ideal body weight.
Patients were not given any advice on their eating,
smoking, or exercise habits during the study.
Administration of any previous antihypertensive
medication was discontinued at least 4 weeks before
the study. The study protocol was approved by the
Hospital Ethics Committee and each patient gave
written consent to participate in the study.
Study protocol is shown in Figure 1. Specifically,
after the 4 weeks’ wash-out period, all patients were
given perindropril (4 mg once daily) for 4 weeks.
Taking into account that the seated diastolic BP after
 Combination therapy with perindopril and diuretics
MS Elisaf et al
788
Figure 1 Study protocol. P: perindopril (4 mg/daily), HCT: hydrochlorothiazide (25 mg/daily), IND: indapamide (2.5 mg/daily).
perindopril treatment was more than 90 mm Hg,
indapamide (2.5 mg once daily) or hydrochlorothia-
zide (25 mg once daily) was added for 4 weeks.
Selection of the diuretic agent was random. Follow-
ing a 4-week wash-out period from the diuretic, in
which only perindopril was given, the alternative
diuretic was administered for another period of 4
weeks. At every visit (every 4 weeks) each patient’s
BP was measured in the right arm with a standard
mercury sphygmomanometer after the patient
remained seated for 10 min. Additionally, heart rate
and body weight were also measured. Compliance
with drug treatment was assessed in each patient by
an interview and a pill count.
In all study periods venous blood was obtained
after a 12 h fast for the determination of serum meta-
bolic parameters (glucose, creatinine, uric acid,
sodium, potassium, magnesium, total cholesterol,
high-density lipoprotein (HDL) cholesterol, low-
density lipoprotein (LDL) cholesterol, apolipopro-
tein A1, apolipoprotein B, and triglycerides).
At the same time, glucose tolerance tests after
ingestion of 75 g of glucose was performed and
samples for serum glucose and insulin were drawn
at 0, and 2 h. In all patients plasma renin activity
(PRA) was measured at the beginning of the study
along with 24-h urine sodium excretion.
Serum creatinine was measured by the method of
Jaffe, serum glucose by the hexokinase method,
serum sodium and potassium by ion selective elec-
trodes, serum magnesium by photometric colori-
metric assay, serum uric acid by the uricase/PAP
method, and serum insulin by the Abbott IMx insu-
lin assay, which is a microparticle enzyme im-
munoassay. Serum cholesterol and triglycerides
were determined by enzymatic colorimetric assay
using an RA1000 analyser (Technicon Instruments,
NY, USA), while HDL cholesterol was determined
enzymatically in the supernatant after precipitation
of other lipoproteins with dextran sulfate-magnes-
ium. LDL cholesterol was calculated using the Frie-
dewald formula. Serum Apo A1 and ApoB were
measured by immunonephelometry with the aid of a
Beckman array analyser (Beckman Instruments, CA,
USA). PRA was measured by a radioimmunoassay.
Statistical analysis
Data are presented as means ± standard deviation.
Blood pressure and biochemical parameters at the
end of each study period were used for statistical
analysis according to that recommended by Hills
and Armitage.14 This method allows comparison of
the effects of combination of perindopril either with
hydrochlorothiazide or with indapamide to be
adjusted for any period effects. This method also
includes a test for treatment-period interaction.
Comparison of the effects of perindopril used either
alone or in combination with diuretics was derived
from the differences among responses for the study
periods. Significance was assessed with Student’s t-
statistic. No carryover effect was detected for any
variable measured and therefore there was no
necessity to analyse any variables as a parallel
study. All statistical tests were two-tailed and were
interpreted at the 5% significance level.
Results
A significant decrease in the mean BP was observed
after perindopril administration in the whole group
of patients [from 118.9 ± 9.3 to 108.9 ± 7.9 mm Hg,
mean reduction of 8.4%, P ⬍ 0.005]. Perindopril
was more efficacious in patients with higher PRA,
since patients with a more than 5 mm Hg decrease
in mean BP (range, 5.8–10.6 mm Hg) had higher PRA
compared to those with a decrease in mean BP less
than 5 mm Hg (range, 0.4 –4.2 mm Hg) [2 ± 0.8
ng/ml/h vs 0.4 ± 0.5 ng/ml/h, P ⬍ 0.01]. Further-
more, there was a good correlation between PRA
values and the decrease in mean BP (r = 0.42, P =
0.02). The addition of either hydrochlorothiazide or
indapamide produced an additional decrease in
mean BP from 108.2 ± 8 mm Hg on monotherapy to
87 ± 9 mm Hg after 4 weeks’ treatment with hy-
drochlorothiazide (mean reduction of 14%, P ⬍
0.005) and from 109 ± 6.8 mm Hg on monotherapy
to 98.5 ± 7 mm Hg after 4 weeks’ treatment with
indapamide (mean reduction of 9.6%, P ⬍ 0.01).
However, the perindopril + hydrochlorothiazide
combination was more effective than the perindopril
+ indapamide combination (P ⬍ 0.05).
No significant changes in serum lipid parameters
were observed after perindopril administration in
the whole group of patients (mean changes of all
lipid parameters ⬍ 5%, P = NS). As shown in Table
1, both diuretics increased the total and LDL choles-
terol levels moderately, though not significantly.
However, there were no significant differences in
the changes of all lipid parameters measured
 Combination therapy with perindopril and diuretics
MS Elisaf et al

789
Table 1 Mean changes (%) in serum lipid parameters after postload glucose levels (from 90 ± 11 mg/dl to
diuretics administration 103 ± 9 mg/dl, P ⬍ 0.01 and from 115 ± 16 mg/dl to
132 ± 12 mg/dl, P ⬍ 0.01, respectively after hydroch-
Parameters After HCT After IND Pa
(mg/dl) (%) (%) lorothiazide administration and from 89 ± 14 mg/dl
to 99 ± 12 mg/dl, P ⬍ 0.01, and from 119 ± 16 mg/dl
Total cholesterol 9.5 7 NS to 132 ± 12 mg/dl, P ⬍ 0.01, respectively after inda-
HDL cholesterol −1.2 −1 NS pamide administration), while there were no sig-
LDL cholesterol 8.5 7 NS nificant differences in the changes of carbohydrate
Triglycerides 6.0 5 NS metabolism parameters between the two combi-
Apolipoprotein A1 −3.0 −2 NS
Apolipoprotein B 9.0 6 NS nation therapies. Even though the fasting and 2 h
postload serum insulin levels were somewhat
a
P between the two combination therapies, HCT: hydrochloro- increased after the addition of both diuretics, the
thiazide, IND: indapamide. changes were not statistically significant (Table 3).

between the two combination therapies. Perindopril Discussion

administration was followed by an increase in
serum potassium levels (from 4.1 ± 0.2 mm Hg to
4.3 ± 0.3 mm Hg, P = 0.08). Both hydrochlorothiaz-
ide and indapamide produced a decrease in serum
potassium levels to lower than pretreatment values
(from 4.4 ± 0.2 mmol/L to 3.95 ± 0.2 mmol/L after
hydrochlorothiazide treatment, and from
4.3 ± 0.15 mmol/L to 4 ± 0.2 mmol/L after indapam-
ide treatment), even though the changes in serum
potassium levels were only marginally statistically
significant (P = 0.08 for hydrochlorothiazide and P
= 0.09 for indapamide) possibly because of the rela-
tively small number of patients studied.
An approximately 10% decrease in serum mag-
nesium levels was found after both diuretics’ admin-
istration. However, there were no significant differ-
ences in the changes of serum electrolytes between
the two combination therapies (Table 2). A signifi-
cant increase in serum uric acid levels was observed
after hydrochlorothiazide administration (from
5.2 ± 1.3 mg/dl to 6 ± 1.4 mg/dl, P ⬍ 0.01), while
indapamide did produce a small insignificant
increase in uric acid levels (from 5.3 ± 0.9 mg/dl to
5.7 ± 0.8 mg/dl, P = NS). Subsequently, the addition
of hydrochlorothiazide caused a greater increase in
serum uric acid levels compared to that observed
after indapamide administration (P ⬍ 0.01). Perin-
dopril treatment was not followed by any change in
either fasting or 2 h postload serum glucose levels
(from 96 ± 12 mg/dl to 91 ± 9 mg/dl, and from
120 ± 12 mg/dl to 118 ± 24 mg/dl, respectively) and
insulin levels (from 10.2 ± 3 mU/L to 9.9 ± 4 mU/L
and from 13 ± 3 mU/L to 12 ± 6 mU/L, respectively).
Both diuretics increased significantly fasting and 2 h
Our study reconfirmed that perindopril is an effec-
tive antihypertensive drug without any adverse
metabolic effects on carbohydrate or lipid metabo-
lism.15 In agreement with previously published data
the drug’s antihypertensive efficacy was related to
the renin-angiotensin axis activation; that is, the BP
fall was well correlated to the pretreatment PRA.16,17
However, a considerable proportion of hyperten-
sive patients exhibit an inadequate response to mon-
otherapy with ACE inhibitors, as was the case in our
patients.3,18 In such cases, the addition of a diuretic
to a standard dose of ACE inhibitor gives a better
response than increasing the dose of ACE inhibitor.4
Thus, the combination of ACE inhibitors with
diuretics appears to represent an ideal choice in
terms of efficacy, compliance, side effects and
cost.19–21 In fact, in our patients a near normalisation
of BP values after diuretics administration was achi-
eved. The synergistic action of ACE inhibitor-
diuretic combination allows the diuretic dose to be
reduced, leading to a reduction in the metabolic side
effects commonly observed with diuretic ther-
apy.8,22,23 Accordingly, previous studies have shown
that the addition of ACE inhibitors to thiazides can
blunt or even prevent the metabolic consequences
of diuretic therapy, while they enhance their antihy-
pertensive effect.6–11 Nonetheless, there seem to be
a few studies concerning the metabolic conse-
quences of added thiazide diuretics to ACE inhibi-
tors. Additionally, comparison of the observed bio-
chemical derangements between hydrochloro-
thiazide and indapamide when added to ACE
inhibitors is lacking. This is of special importance,
since indapamide is an effective and safe antihyper-

Table 2 Mean changes (%) in serum metabolic parameters after

diuretics administration Table 3 Mean changes (%) of carbohydrate metabolism para-
meters after diuretics administration
Serum parameters After HCT After IND Pa
administration administration Parameters After HCT After IND Pa
(%) (%) administration administration
(%) (%)
Potassium (mmol/L) −10.5 −8.4 NS
Sodium (mmol/L) −2.0 −1.5 NS Fasting glucose (mg/dl) 14* 11* NS
Magnesium (mmol/L) −9.0 −8.0 NS 2 h glucose (mg/dl) 14.5* 10.5* NS
Creatinine (mg/dl) 7.0 5.0 NS Fasting insulin (mU/L) 11 9 NS
Uric acid (mg/dl) 15.5* 7.5 0.05 2 h insulin (mU/L) 7 6 NS

a a
P between the two combination therapies. Between the two combination therapies.
*P ⬍ 0.01 compared to values obtained during P treatment. *P ⬍ 0.01 compared to values obtained during P treatment.
HCT: hydrochlorothiazide, IND: indapamide, P: perindopril. HCT: hydrochlorothiazide, IND: indapamide, P: perindopril.
 Combination therapy with perindopril and diuretics
MS Elisaf et al
790
tensive drug, thought to be unique among diuretics
devoid of lipid or other metabolic effects.12,13 How-
ever, our study showed that hydrochlorothiazide at
a dose of 25 mg/day was more effective than indapa-
mide at a dose of 2.5 mg/day, when both drugs are
added in patients on perindopril treatment. Thus,
our results point to the suggestion that a hydrochlor-
othiazide dose of not less than 25 mg is necessitated
to achieve optimal BP lowering.7,9 Nevertheless, at
this dosage a number of adverse metabolic effects
was observed, namely a 10% increase in serum total
and LDL cholesterol levels, an increase in serum
uric acid levels, and, most importantly, a significant
increase in serum glucose levels. Similar changes,
however, even though of lower degree, were also
found after indapamide treatment, while the only
significant difference between the two combination
therapies concerns the increase in serum uric acid
levels, which was significantly higher after hydroch-
lorothiazide administration. Interestingly, after both
diuretics’ treatment a significant increase in fasting
and 2 h postload glucose levels was evident, while
no significant change in both fasting and 2 h post-
load insulin levels was found, suggesting that
diuretics could directly or indirectly (through
hypokalaemia) decrease pancreatic insulin
secretion.24,25
Even though evaluation of insulin sensitivity
using the euglycaemic glucose clamp technique was
not performed in our study, it is noteworthy that no
beneficial effect of perindopril on serum glucose and
insulin levels both fasting and 2 h postloading rela-
tive to baseline was demonstrated. Thus, the puta-
tive beneficial effect of ACE inhibitors on carbo-
hydrate metabolism remains controversial.26–29 It is
possible that the impact on sensitivity to insulin var-
ies with different ACE inhibitors,30 and it is depen-
dent on the degree of underlying insulin resistance.
If subjects are not particularly insulin resistant, as
was the case in our patients, then it may be more
difficult to reveal a beneficial effect.26 Furthermore,
the ACE inhibitor dosage could have also played a
role in its influence in the carbohydrate tolerance,
since in one study the administration of only high
doses of captopril (⬎100 mg/day) mitigated the
hyperglycaemic effect of hydrochlorothiazide.6
Therefore, a placebo-controlled study to address the
effect of ACE inhibition on sensitivity to insulin in
essential hypertension needs to be carried out.
A small increase in serum potassium levels was
observed after perindopril therapy and a decrease to
lower than pretreatment values followed the two
combination therapies. Nevertheless, the 25 mg of
hydrochlorothiazide did not induce a significant
decrease in serum potassium levels to hypokalaemic
levels, suggesting that perindopril guards against the
hypokalaemic effects of the added diuretics. The
changes in potassium homeostasis could have
played a role in the diuretic-induced adverse effects
on carbohydrate metabolism. It has been suggested
that the decrease in serum potassium levels might
impair insulin secretion and inhibit muscle glu-
cogen synthase activity, which is associated with a
decrease in insulin-mediated uptake of glucose.24,31
Our data clearly showed that the administration
of ACE inhibitor did not ameliorate the adverse
effects of diuretics in agreement with recently pub-
lished data, which reported that combination ther-
apy evoked hepatic insulin resistance when 5 mg of
bendrofluazide was used.26 Consequently, great care
needs to be exercised in the use of these combi-
nations notwithstanding that they are particularly
efficacious in lowering BP, as it is also illustrated in
the present study. It should be mentioned that it is
not certain what the effects of ACE inhibitors com-
bined with lower doses of hydrochlorothiazide will
be. This is of special importance, since many of the
combined proprietary preparations of ACE inhibi-
tors and thiazides available in Europe contain
smaller quantities of thiazide diuretics compared to
those used in this study.
References
1 Berne C. Metabolic effects of ACE inhibitors. J Intern
Med 1991; 229 (Suppl 2): 119–125.
2 Townsend RR, Bryan Holland O. Combination of con-
verting enzyme inhibitor with diuretic for the treat-
ment of hypertension. Arch Intern Med 1990; 150:
1175–1183.
3 McAreavey D et al. ‘Third Drug’ trial: comparative
study of antihypertensive agents added to treatment
when blood pressure remains uncontrolled by a beta-
blocker plus thiazide diuretic. Br Med J 1984; 288:
106–110.
4 Sassano P et al. Comparison of increase in the enalap-
ril dose and addition of hydrochlorothiazide as
second-step treatment of hypertensive patients not
controlled by enalapril alone. J Cardiovasc Pharmacol
1989; 13: 314 –319.
5 Weinberger MH. Comparison of captopril and hyd-
rochlorothiazide alone and in combination in mild to
moderate essential hypertension. Br J Clin Pharmacol
1982; 14: 127S–131S.
6 Weinberger MH. Influence of an angiotensin con-
verting-enzyme inhibitor on diuretic-induced meta-
bolic effects in hypertension. J Hypertens 1983; 5
(Suppl 3): 132–138.
7 Scholze J, Group for the East Germany Collaborative
Trial. Short report: ramipril and hydrochlorothiazide
combination therapy in hypertension: a clinical trial
of factorial design. J Hypertens 1993; 11: 217–221.
8 Chrysant SG, The Lisinopril–Hydrochlorothiazide
Group. Antihypertensive effectiveness of low-dose
hydrochlorothiazide combination. A large multicenter
study. Arch Intern Med 1994; 154: 737–743.
9 Canter D et al. Quinapril and hydrochlorothiazide
combination for control of hypertension: assessment
by factorial design. J Hum Hypertens 1994; 8: 155–162.
10 Lacourcière Y, Gagnè C. Influence of combination of
captopril and hydrochlorothiazide on plasma lipids,
lipoproteins and apolipoproteins in primary hyperten-
sion. J Hum Hypertens 1993; 7: 149–152.
11 Weinberger MH. Influence of an angiotensin-con-
verting enzyme inhibitor on diuretic-induced meta-
bolic effects in hypertension. Hypertension 1983; 5:
132–138.
12 Osei K, Holland G, Falko M. Indapamide. Effects on
apoprotein, lipoprotein, and glucoregulation in ambu-
latory diabetic patients. Arch Intern Med 1986; 146:
1973–1977.
13 Campbell DB, Moore RA. The pharmacology and clini-
cal pharmacology of indapamide. Postgrad Med J 1981;
57 (Suppl 2): 7–17

14 Hills M, Armitage P. The two-tailed cross-over trial. Br
J Clin Pharmacol 1979; 8: 7–10.
15 Lees KR et al. Captopril versus perindopril: a double-
blind study in essential hypertension. J Hum Hyper-
tens 1989; 3: 17–22.
16 Case DB et al. Possible role of renin in hypertension
as suggested by renin-sodium profiling and inhibition
of converting enzyme. N Engl J Med 1977; 296: 641–
646.
17 MacGregor GA et al. Captopril in essential hyperten-
sion; contrasting effects of adding hydrochlorothiazide
or propranolol. Br Med J 1982; 284: 693–696.
18 Veterans Administration Cooperative Study Group on
Antihypertensive Agents. Captopril: evaluation of low
doses, twice-daily doses and the addition of diuretic
for the treatment of mild to moderate hypertension.
Clin Sci 1982; 63 (Suppl 8): 443S– 445S.
19 Holland OB, Kuhnert LV, Campbell WB, Anderson RJ.
Synergistic effect of captopril with hydrochlorothiaz-
ide for the treatment of low-renin hypertensive black
patients. Hypertension 1983; 5: 235–239.
20 Atkinson AB, Brown JJ, Lever AF, Robertson JIS. Com-
bined treatment of severe intractable hypertension
with captopril and diuretic. Lancet 1980; ii: 105–108.
21 Weinberger MH. Blood pressure and metabolic
responses to hydrochlorothiazide, captopril, and the
combination in black and white mild-to-moderate
hypertensive patients. J Cardiovasc Pharmacol 1985;
7: S52–S55.
22 Dahlöf B et al. Potentiation of the antihypertensive
effect of enalapril by randomised addition of different
doses of hydrochlorothiazide. J Hypertens 1985; 3:
S483–S486.
23 Andren L, Weiner L, Svensson A, Hanson L. Enalapril
with either a “very low” or “low” dose of hydrochloro-
Combination therapy with perindopril and diuretics
MS Elisaf et al
791
thiazide is equally effective in essential hypertension.
A double-blind trial in 100 hypertensive patients.
J Hypertens 1983; 1: 384 –386.
24 Fajan SS et al. Benzothiazidine suppression of insulin
release from normal and abnormal islet cell tissue in
man. J Clin Invest 1966; 45: 481– 493.
25 Santoro D et al. Effect of chronic angiotensin-con-
verting enzyme inhibition on glucose tolerance and
insulin sensitivity in essential hypertension. Hyper-
tension 1992; 20: 181–191.
26 Hunter SJ et al. Effects of combination therapy with an
angiotensin converting enzyme inhibitor and thiazide
diuretic on insulin action in essential hypertension.
J Hypertens 1998; 16: 103–109.
27 Pollare T, Lithell H, Selinus I, Berne C. A comparison
of the effects of hydrochlorothiazide and captopril on
glucose and lipid metabolism in patients with hyper-
tension. N Engl J Med 1989; 321: 868–873.
28 Seghieri G et al. Effect of chronic ACE inhibition on
glucose tolerance and insulin sensitivity in hyperten-
sive type 2 diabetics. Diabetic Med 1992; 9: 732–738.
29 Prince MJ et al. Metabolic effects of hydrochlorothiaz-
ide and enalapril during treatment of the hypertensive
diabetic patient. Arch Intern Med 1988; 148: 2363–
2368.
30 Uehara M et al. Effect on insulin sensitivity of angiot-
ensin converting enzyme inhibitors with or without a
sulphydryl group: bradykinin may improve insulin
resistance in dogs and humans. Diabetologia 1994; 37:
300–307.
31 Bogardus C, Lillioja S, Stone K, Mott D. Relationship
between muscle glycogen synthase activity on in vivo
insulin action in man. J Clin Invest 1984; 73: 1185–
1190.